Seismic vulnerability of RC frame and shear wall structures in singapore

This study focuses on seismic vulnerability of frame and shear wall structures in Singapore, designed primarily for gravity loads, when they are subjected to far field effects of earthqu

SHEAR WALL STRUCTURES IN SINGAPORE

LI ZHIJUN

NATIONAL UNIVERSITY OF SINGAPORE

2006

LI ZHIJUN

(M.ENG., B.ENG., SOUTH CHINA UNIVERSITY OF TECHNOLOGY)

A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF CIVIL ENGINEERING

NATIONAL UNIVERSITY OF SINGAPORE

2006

ACKNOWLEDGEMENT

I would like to take this opportunity to express my profound gratitude and

sincere appreciation to my supervisor Professor T Balendra and my co-supervisor

Associate Professor Tan Kiang Hwee, for their kind guidance, systematic guidance and

supervision throughout the course of this study

I also like to thank the staffs of the Structural Laboratory for their help and

advice Many thanks to Mr Sit Beng Chiat, Mr Edgar Lim, Mr Ang Beng Onn, Ms

Annie Tan, Mr Ow Weng Moon, Mr Kamsan Bin Rasman, Mr Yip Kwok Keong, Mr

Ong Teng Chew, Mr Yong Tat Fah, Mr Wong Kah Wai, Stanley, and Mr Martin who

help in many ways in the experiment Special acknowledgement is given to Mr Choo

Peng Kin, Mr Koh Yian Kheng and Mr Ishak Bin A Rahman who had assisted and

guided me tremendously in the experiment

Gratitude is extended to my seniors Dr Kong Kian Hau, Dr Kong Sia Keong

and Ms Suyanthi Sakthivel; friends and colleagues Mr Michael Perry, Ms Wu Hong,

Mr Duan Wen Hui, Mr Zhou En Hua, Ms Yu Hongxia, Mr Kan Jian Han, Mr Wiryi

Aripin, Mr Chen Jun, Mr Zhao Dian Feng, Mr Gao Xiao Yu, Dr Li Jin Jun and Ms

Zhou Yu Qian for their help and encouragement

I am greatly indebted to my mother and brother who have encouraged me a lot

and made many sacrifices during the study

I am grateful to my lecturers, relatives and friends who have supported the study

in many ways

LIST OF SYMBOLS xvi

CHAPTER 1 INTRODUCTION 1

1.1 B ACKGROUND 1

1.2 L ITERATURE REVIEW 3

1.2.1 Overview of seismic studies of RC GLD structures 3

1.2.2 Research of GLD buildings designed according to ACI code 5

1.2.3 Research of GLD buildings designed according to Korean nonseismic detailing 8

1.2.4 Research of GLD buildings in Singapore designed according to BS8110 code 9

1.2.5 Seismic demand and seismic adequacy evaluation for buildings in Singapore 17

1.2.6 Overview of seismic retrofitting of GLD buildings 20

1.3 O BJECTIVE AND SCOPE 23

1.4 O RGANIZATION OF THE THESIS 24

CHAPTER 2 EXPERIMENTAL STUDY OF A 4-story FRAME STRUCTURE 29

2.1 I NTRODUCTION 29

2.2 E XPERIMENTAL MODEL 30

2.2.1 Model scaling similitude 31

2.2.2 Material properties 32

2.3 T EST SETUP AND TEST PROCEDURE 34

2.3.1 Details of the setup 34

2.3.2 Instrumentation 36

2.3.3 Loading history and test procedure 37

2.4 E XPERIMENT RESULTS AND INTERPRETATION 38

2.4.1 Global response 38

2.4.2 Local responses 45

2.4.3 Moment-curvature curves of the sections 47

2.5 S UMMARY 50

CHAPTER 3 DEVELOPMENT OF THE FEA MODEL FOR FRAMES 70

3.1 I NTRODUCTION 70

3.2 FEA MODEL USING RUAUMOKO 70

3.2.1 Overview of RUAUMOKO 70

3.2.2 FEA modeling 72

3.3 C OMPARISON OF FEA AND EXPERIMENTAL RESULTS 77

3.3.1 Natural periods 77

3.3.2 Load-displacement curves 78

3.3.3 Failure mode 79

3.3.4 FEA model for the full scale structure 80

3.4 S UMMARY 81

CHAPTER 4 EXPERIMENTAL STUDY OF A 25-story SHEAR WALL STRUCTURE 89

4.1 I NTRODUCTION 89

4.2 E XPERIMENTAL MODEL 90

4.2.1 Scale factor 91

4.2.2 Material scaling simulation 93

4.2.3 Material properties 95

4.3 T EST SETUP AND TEST PROCEDURE 97

4.3.1 Details of the setup 97

4.3.2 Instrumentation 99

4.3.3 Loading history and test procedure 100

4.4 E XPERIMENTAL RESULTS AND INTERPRETATION 101

4.4.1 Global response 101

4.4.2 Local response 108

4.5 S UMMARY 111

CHAPTER 5 DEVELOPMENT OF THE FEA MODELS FOR SHEAR WALLS 150

5.1 I NTRODUCTION 150

5.2 FEA MODELS USING RUAUMOKO 150

5.2.1 2D FEA modeling 151

5.2.2 3D FEA modeling 154

5.2.3 Comparison of FEA results using RUAUMOKO with experimental results 159

5.3 FEA MODELING USING ABAQUS 160

5.3.1 FEA modeling of the control specimen (S1) test 161

5.3.2 FEA modeling of the FRP wrapped specimen (S2) 163

5.3.3 Parameters to identify failure in FEA study 165

5.3.4 Correlation of FEA and experimental results 167

5.4 S UMMARY 170

CHAPTER 6 SEISMIC DEMAND AND CAPACITY 191

6.1 I NTRODUCTION 191

6.2 S EISMIC DEMAND 192

6.2.1 Accelerations and response spectra of two recent strong earthquakes 192

6.2.2 Maximum possible earthquake that could affect Singapore 194

6.2.3 Selected sites 196

6.2.4 Surface motions and amplification factors 197

6.3 M ETHODS OF ANALYSIS AND FAILURE IDENTIFICATION 199

6.3.1 Methods of analysis 199

6.3.2 Failure identification 202

6.4 C ASE STUDY 1: A 25- STORY REINFORCED CONCRETE HDB POINT BLOCK 204

6.4.1 FEA modeling 204

7.1 C ONCLUSIONS 257

7.2 R ECOMMENDATIONS 259

REFERENCES 260

APPENDIX A CALCULATION OF PARAMETERS FOR RUAUMOKO (2D VERSION) 268

A.2.1 Parameters needed to be defined 269

A.2.2 Determination of the parameters 270

APPENDIX B CALCULATION OF SHEAR FORCE CAPACITY 284

APPENDIX C CALCULATION OF PARAMETERS FOR RUAUMOKO (3D VERSION) 285

C.1 E LASTIC SECTION PROPERTIES 285

C.2 P ARAMETERS FOR THE AXIAL FORCE - MOMENT INTERACTION YIELD SURFACE 286

C.3 P ARAMETERS FOR BEAM FLEXURAL YIELD CONDITIONS 288

APPENDIX D PROCEDURE FOR CALCULATION OF RESPONSE SPECTRA 289

APPENDIX E BEDROCK ACCELEROGRAMS FOR THE DESIGN EARTHQUAKE 296

APPENDIX F INPUT FILES OF SHAKE91 300

F.1 I NPUT FILE FOR THE KAP SITE 300

F.2 I NPUT FILE FOR THE KAT SITE 302

F.3 I NPUT FILE FOR THE MP SITE 304

APPENDIX G IDENTIFICATION OF GLOBAL FLEXURAL FAILURE 306

APPENDIX H SECTIONAL PROPERTIES OF FEA MODELS IN CASE STUDIES 307

H.1 C ASE STUDY 1 : A 25- STORY REINFORCED CONCRETE POINT BLOCK 307

H.2 C ASE STUDY 2 : A SUB - FRAME OF A 4- STORY FRAME BUILDING 312

SUMMARY

Because Singapore is located on a stable part of the Eurasian Plate, with the

nearest earthquake fault 400 km away in Sumatra, buildings in Singapore were

designed according to the British Standard without any seismic provision However,

due to the far-field effects of earthquakes in Sumatra (Balendra et al 1990), they are

occasionally subjected to tremors due to earthquakes occuring at the Sumatra In the

last two years (2004 and 2005), tremors were felt five times in Singapore due to the

strong earthquakes at Sumatra, which highlight the earthquake threat to Singapore

This study focuses on seismic vulnerability of frame and shear wall structures in

Singapore, designed primarily for gravity loads, when they are subjected to far field

effects of earthquakes in Sumatra The evaluation of the seismic vulnerability is

achieved by comparing the demand curve and capacity curve in the acceleration-

displacement (A-D) format

The demand curve is obtained based on the accelerograms of bedrock motions

due to the worst earthquake scenario in Sumatra, and soil profiles of the selected sites

(located at Marine Parade, Katong Park and Katong area) The worst earthquake

scenario is identified as an earthquake with Mw=9.5, at 600 km away from Singapore,

by incorporating the data from two recent earthquakes that occurred in Sumatra

March 28 2005)

To establish the accuracy by FEA analytical model to determine the capacity of

a full scale building, experimental studies of a 1/5-scale shear wall model and a

the cyclic test, by comparing the results from the pushover tests with those from the

cyclic tests In the frame tests, a strong column–weak beam mechanism was observed,

although the frame was designed according to BS8110(1985) without any seismic

provision And the results from the shear wall tests revealed that the shear walls fail at

the base due to shear Retrofitting using glass fiber reinforced polymer (GFRP) system

was proposed, and the cyclic behavior of shear wall structures retrofitted with GFRP

system was investigated experimentally The FEA model for the GFRP retrofitted

structure was established and validated using the test results

Two case studies have been carried out for the vulnerability study: (1) a 4-story

frame building, representing typical low-rise buildings; and (2) a 25-story shear

wall-frame building, representing typical high-rise buildings In the case studies, the

pushover and dynamic collapse analysis for the full scale structures are carried out

From these two case studies, it is concluded that low-rise buildings in Singapore would

meet the demand, but in certain cases, high-rise buildings in Singapore may suffer

some damage due to the worst possible earthquake For such insufficient cases, a

seismic retrofitting scheme using FRP system is proposed

LIST OF FIGUES

Figure 1.1 Sumatra fault and subduction of the Indian-Australian Plate into Eurasian

Plate (Balendra et al 2001) 27

Figure 1.2 Typical load-displacement relationship for a reinforced concrete member (Paulay and Priestley 1992) 27

Figure 1.3 Modified Takeda Hysteresis 28

Figure 2.1 Prototype structure: (a) plan view of the whole building (b) selected critical frame (c) two story- one and a half bay frame chosen for the test model 52

Figure 2.2 3D view of the test frame specimen 52

Figure 2.3 The experimental model: (a) test specimen dimension (b) reinforcement details in columns (c) cross section of columns (d) reinforcement details in beams 53

Figure 2.4 The stress- strain curve of steel reinforcement used in model 54

Figure 2.5 3D view of the whole frame steel cage 54

Figure 2.6 3D view of the lap splice of columns above the base block 55

Figure 2.7 3D view of the lap splice of columns above the 1st story joints 55

Figure 2.8 3D view of the 2nd story joints 56

Figure 2.9 3D view of the test set-up 56

Figure 2.10 Side view of the set-up 57

Figure 2.11 Details of the lateral whiffle tree loading system 58

Figure 2.12 3D view of the lateral loading whiffle tree system 59

Figure 2.13 3D view of the lateral support 59

Figure 2.14 Two jacks were used together for one column 60

Figure 2.15 Locations of the strain gauges on the reinforcing bars 60

Figure 2.16 Locations of the strain gauges on the concrete surface 61

Figure 2.17 Locations of transducers 61

Figure 2.18 3D view of the transducers at the 1st story external joint 62

Figure 2.19 3D view of the omega gauges used at a joint 62

Figure 2.20 Cyclic loading history 63

Figure 2.21 Crack pattern and failure mode of specimen S1 (a) front view (b) back view 63

Figure 2.22 Breaking of the outermost tension reinforcing bars: (a) location of the base column (b) location of the beam-column interface 64

Figure 2.23 Crack pattern and failure mode of specimen S2 (a) front view (b) back view 64

Figure 2.24 Load-displacement relationship: (a) 2nd floor displacement (b) 1st floor displacement 65

Figure 2.25 Joint rotation histories in pushover test (a) 1st story joints (b) 2nd story joints 66

Figure 2.26 Base shear force (kN) vs curvature (rad/mm) curves at different locations 67

Figure 2.27 Moment-curvature curves in the pushover test 68

RUAUMOKO: (a) moment envelope; (b) shear envelope 88

Figure 3.7 Comparison of FEA results with individual stiffness reduction factors and with average stiffness reduction factors 88

Figure 4.1 Plan view of 25-story point block 114

Figure 4.2 Plan view of prototype wall (a) dimensions (b) identification of segments 115

Figure 4.3 3D view of the specimens (a) control specimen (b) FRP wrapped specimen 116

Figure 4.4 Plan view and geometry of the test model 117

Figure 4.5 Overall 3D view of the rebar in the wall specimen 117

Figure 4.6 Plan view of the reinforcing bar geometry 118

Figure 4.7 Details of reinforcing bars in the base block reinforcing bars 119

Figure 4.8 Average stress- strain curves of steel reinforcement used in model 120

Figure 4.9 Concrete casting in the lab 120

Figure 4.10 Wall after the application of MBT primer (Note the rounded edge of the wall) 121

Figure 4.11 Locations of FRP bolts (front view) 122

Figure 4.12 Locations of FRP bolts (side view) 123

Figure 4.13 3D view of the overall test setup for the control wall (specimen S1) 124 Figure 4.14 3D view of test setup for FRP wrapped wall (specimen S2) 124

Figure 4.15 Front view of the overall set-up 125

Figure 4.16 Side view of the overall set-up 126

Figure 4.17 Plane view of the loading system 127

Figure 4.18 3D view of connections of actuator to P beam and P beam to U beams 128

Figure 4.19 3D view of connections of U beams to L angles and L angles to walls 128

Figure 4.20 3D view of post-tension strands anchored to the U beams 129

Figure 4.21 3D view of the lateral supporting system 129

Figure 4.22 Locations of strain gauges on the reinforcing bars (a) left flange wall (b) web wall 130

Figure 4.23 Locations of strain gauges on concrete (a) left flange wall (b) web wall 131

Figure 4.24 Strain gauges on the FRP of the wall (a) left flange wall (b) web wall (c) right flange wall 132

Figure 4.25 Locations and labels of displacement transducers (range of the

displacement transducer is indicated within brackets) 133

Figure 4.26 Cyclic loading history 134

Figure 4.27 3D view of the wall model after white wash 134

Figure 4.28 3D view of crack pattern of the flange wall 135

Figure 4.29 First spalling of the concrete of the right flange wall 135

Figure 4.30 Overview of the shear failure mode of the right flange wall 136

Figure 4.31 Spalling concrete of the upper right part of the right flange wall 136

Figure 4.32 Spalling of concrete of the bottom part of the right flange wall 137

Figure 4.33 Shear cracks on the left flange wall 137

Figure 4.34 Spalling of the left flange wall corner 138

Figure 4.35 First FRP debonding of the left flange wall 138

Figure 4.36 First FRP debonding of the right flange wall 139

Figure 4.37 First FRP debonding of the web wall 139

Figure 4.38 FRP debonding of the right flange wall (the second 15 mm cycle) 140

Figure 4.39 Crushing of the corner of the right flange wall at 16 mm top displacement 140

Figure 4.40 Debonding of FRP at the right flange wall corner 141

Figure 4.41 The left flange wall concrete crushing 141

Figure 4.42 Crushing of the right flange wall corner at 24 mm top displacement 142 Figure 4.43 Load-displacement relationships at the top actuator level and the 1st floor level of the flange wall (control specimen S1) 142

Figure 4.44 Load-displacement relationships of flange wall and web wall at the 1st floor level (control specimen S1) 143

Figure 4.45 Load-displacement relationships of the top actuator level and the 2nd floor level of the flange wall (FRP wrapped specimen S2) 143

Figure 4.46 Load-displacement relationships at the 2nd and the 1st floor level of the flange wall (FRP wrapped specimen S2) 144

Figure 4.47 Load-displacement relationship of the FRP wrapped specimen at the 1st floor level of the flange wall and the web wall 144

Figure 4.48 Comparison of the load- top actuator level displacement relationships between cyclic loading and pushover loading for the non-FRP specimen 145

Figure 4.49 Comparison of the load-top actuator level displacement relationships between cyclic loading and pushover loading for the FRP wrapped specimen 145

Figure 4.50 Comparison of the load-top actuator level displacement relationships between FRP wrapped specimen and control specimen under cyclic loading 146

Figure 4.51 Comparison of the load-top actuator level displacement relationships between FRP wrapped specimen and control specimen under cyclic loading (Cycle by Cycle) 147

Figure 4.52 Load (kN) vs strain in reinforcing bars (micro strain) curves at different locations (Specimen S1) 148

(elevate view) 175

Figure 5.3 Nodes, elements and sectional properties of the 3D FEA modeling 176

Figure 5.4 The frame element in RUAUMOKO 3D versionc(Carr 2002b) 176

Figure 5.5 Comparison of results for specimen S1 between the pushover FEA using RUAUMOKO (2D and 3D) and the cyclic test 177

Figure 5.6 Comparison of results for specimen S1 between the test and the FEA using RUAUMOKO (2D pushover and cyclic analysis) 177

Figure 5.7 Cycle by cycle comparison between the test and 2D cyclic FEA using RUAUMOKO for specimen S1 178

Figure 5.8 Comparison of results for FRP retrofitted specimen S2 between the pushover FEA using RUAUMOKO (2D and 3D) and the cyclic test 179

Figure 5.9 Comparison of results for specimen S2 between the test and the FEA using RUAUMOKO (2D pushover and cyclic analysis) 179

Figure 5.10 Cycle by cycle comparison between the test and 2D cyclic FEA using RUAUMOKO for the specimen S2 180

Figure 5.11 3D view of the modeling of the control wall (S1) 181

Figure 5.12 The stress-strain curve for steel used in ABAQUS 181

Figure 5.13 The stress-strain curve of concrete damaged plasticity model used in ABAQUS 182

Figure 5.14 3D view of the modeling of the FRP wrapped specimen (S2) 182

Figure 5.15 The stress-strain curve of the confined concrete (Teng 2001) 183

Figure 5.16 Wall divided into regions 183

Figure 5.17 The stress-strain curves for the confined concrete at different regions 184

Figure 5.18 Initial shear failure in reinforced concrete flange walls of the control specimen (S1) at 54.4 kN 184

Figure 5.19 Final shear failure in the control specimen (S1) at 113.9 kN 185

Figure 5.20 Axial compressive stress in the control specimen (S1) 185

Figure 5.21 Initial shear failure in reinforced concrete flange wall of FRP wrapped specimen (S2) at 53.6kN 186

Figure 5.22 Initial shear failure of FRP wrapped specimen (S2) due to FRP debonding at the 3rd cycle 186

Figure 5.23 Shear failure of FRP wrapped specimen (S2) due to FRP debonding at the 5th cycle 187

Figure 5.24 Shear failure of FRP wrapped specimen (S2) due to FRP rupture (At the

end of the 7th cycle, lateral force was 151.78kN) 187 Figure 5.25 Comparison of pushover FEA using ABAQUS and 3D FEA using

RUAUMOKO without consider the initial stiffness reduction for specimen S1 188 Figure 5.26 Cycle by cycle comparison between experiment and finite element

analysis for Control specimen S1 189 Figure 5.27 Cycle by cycle comparison between experiment and finite element

analysis for FRP wrapped specimen S2 190 Figure 6.1 Locations of the Aceh earthquakes occurred in 26 December 2004 and

the Nias earthquake occurred in 28 March 2005 236 Figure 6.2 Acceleration response spectra of the Aceh earthquake (east-west

direction) 236 Figure 6.3 Acceleration response spectra of the Aceh earthquake (north-south

direction) 237 Figure 6.4 Acceleration response spectra of the Aceh earthquake (vertical direction)

237 Figure 6.5 Velocity response spectra of the Aceh earthquake (east-west direction)

238 Figure 6.6 Velocity response spectra of the Aceh earthquake (north-south direction)

238 Figure 6.7 Velocity response spectra of the Aceh earthquake (vertical direction)

239 Figure 6.8 Acceleration response spectra of the Nias earthquake (east-west direction)

239 Figure 6.9 Acceleration response spectra of the Nias earthquake (north-south

direction) 240 Figure 6.10 Acceleration response spectra of the Nias earthquake (vertical direction)

240 Figure 6.11 Velocity response spectra of the Nias earthquake (east-west direction)

241 Figure 6.12 Velocity response spectra of the Nias earthquake (north-south direction)

241 Figure 6.13 Velocity response spectra of the Nias earthquake (vertical direction) 242 Figure 6.14 Average acceleration response spectra of the design earthquake at

bedrock (Mw=9.5, 600 km away) 242

Figure 6.15 Shear modulus/ shear modulus at low strain 0.001% (G/Gmax) vs shear

strain (%) for clay and sand 243 Figure 6.16 Soil damping ratio vs shear strain(%) for clay and sand 243 Figure 6.17 One of the twelve surface accelerograms of MP site due to design

earthquake at bedrock 244 Figure 6.18 One of the twelve surface accelerograms of KAP site due to design

earthquake at bedrock 244

KAT sites for structural damping ratio of 5% due to design earthquake at bedrock 246 Figure 6.23 Typical story model layout of the 25-story building (plan view) 247 Figure 6.24 3D view of the FEA mesh of the 25-story building 248

Figure 6.25 Loading shape for the pushover analysis (f cu =20 MPa, ultimate loading

case) 248

Figure 6.26 Total base shear demand V b / gravity load W g vs overall drift curve for the

case of f cu =30 MPa, ultimate loading, in x direction (W g = 38866.12kN,

H=64.75m) 249 Figure 6.27 Relationship between V b / W g and shear forces in the critical member (1st

story I-shape flange wall, I3) of the dynamic collapse analysis (f cu =30 MPa, ultimate loading case, in x direction) 249 Figure 6.28 Relationship between scaling factors and shear forces in the critical

member (1st story I-shape flange wall, I3) of the dynamic collapse

analysis (f cu =30 MPa, ultimate loading case, in x direction) 250

Figure 6.29 Total base shear demand V b / gravity load W g vs overall drift curve for the

case of f cu =30 MPa, ultimate loading, in y direction 250

Figure 6.30 Relationship between V b / W g and shear forces in the critical member (1st

story I-shape web wall, I1) of the dynamic collapse analysis (f cu =30 MPa, ultimate loading case, in y direction) 251 Figure 6.31 Relationship between scaling factors and shear forces in the critical

member (1st story I-shape web wall, I1) of the dynamic collapse analysis

(f cu =30 MPa, ultimate loading case, in y direction) 251 Figure 6.32 Seismic capacity curves in x direction obtained from the pushover

adaptive analysis 252 Figure 6.33 Seismic capacity curves in y direction obtained from the pushover

adaptive analysis 252

Figure 6.34 Spectra acceleration (S a )– spectra displacement (S d )curves in x direction

of the 25-story structure (combination of capacity curves and demand curves) 253

Figure 6.35 Spectra acceleration (S a )– spectra displacement (S d ) curves in y direction

of the 25-story structure (combination of capacity curves and demand curves) 253 Figure 6.36 Intercept points of the three insufficient cases 254 Figure 6.37 Elevation view of the 4-story sub-frame (dimension in mm) 254

Figure 6.38 Nodes, elements and sectional properties of FEA for the 4-story

sub-frame 255

Figure 6.39 Total base shear demand V b / gravity load W g vs overall drift curve of the

4-story sub-frame (W g = 957.44kN, H=11.3m) 255 Figure 6.40 Relationship between V b / W g and moment in the critical member

(Element 3) of the dynamic collapse analysis 256 Figure 6.41 Relationship between scaling factors and moment in the critical member

(Element 3) of the dynamic collapse analysis 256

Table 3.1 Values of elastic section properties and bilinear factors (Specimen S1)

83

Table 3.2 Values of elastic section properties and bilinear factors (Specimen S2) 83

Table 3.3 Values to define the yield surface (Specimen S1) 83

Table 3.4 Values to define the yield surface (Specimen S2) 83

Table 3.5 Comparison of maximum moment between the pushover analysis and test 83

Table 3.6 Moment and shear capacities compared with the predicted maximum moment and shear of specimen S1 84

Table 4.1 The comparison of model and prototype I-shape wall 112

Table 4.2 Mbrace EG900 glass fiber reinforced ploymer 113

Table 4.3 Steel reinforcement properties 113

Table 4.4 Parameters of wall specimens tested 113

Table 4.5 Properties of Mbrace primer and saturant 113

Table 5.1 Parameters of elastic section properties and bilinear factor r (specimen S1, 2D FEA) 172

Table 5.2 Parameters of elastic section properties and bilinear factor r (specimen S2, 2D FEA) 172

Table 5.3 Parameters to define the axial load-moment interaction yield surface (specimen S1, 2D FEA) 172

Table 5.4 Parameters to define the axial load-moment interaction yield surface (specimen S2, 2D FEA) 172

Table 5.5 Parameters of elastic section properties and bilinear factor r (specimen S1, 3D FEA) 172

Table 5.6 Parameters of elastic section properties and bilinear factor r (specimen S2, 3D FEA) 173

Table 5.7 Parameters to define the axial load-moment interaction yield surface (specimen S1, 3D FEA) 173

Table 5.8 Parameters to define the axial load-moment interaction yield surface (specimen S2, 3D FEA) 173

Table 5.9 Shear capacity of the shear wall 173

Table 5.10 Comparison between experiment and FEA 174

Table 6.1 Comparison of the motions due to the Aceh earthquake in December 2004 and the Nias earthquake in March 2005 229

Table 6.2 Prediction of peak rock motion 229

Table 6.3 Soil data for the Marine Parade (MP) site 230

Table 6.4 Soil data for the Katong Park (KAP) site 231

Table 6.5 Soil data for the Katong area (KAT) site 231

Table 6.6 The fundamental period obtained from the modal analysis 232

Table 6.7 Gravity loads and lateral forces (1% of the total gravity loads) applied at the story levels (unit: kN ) 232

Table 6.8 RUAUMOKO pushover and dynamic analysis results 233

Table 6.9 Minimum thickness requirement and layers of GFRP sheets for retrofitting (f cu=30 MPa, ultimate loading case, in y direction) 234

Table 6.10 Minimum thickness requirement and layers of GFRP sheets for retrofitting (f cu=30 MPa, common loading case, in y direction) 234

Table 6.11 Minimum thickness requirement and layers of GFRP sheets for retrofitting (f cu=20 MPa, common loading case, in y direction) 235

Table 6.12 Dimensions and reinforcement details of members of the 4-story sub-frame 235

As Effective shear area

Sectional area of tensile reinforcing bars

Reinforcing steel bars stock holders & distributors

fc' Concrete compressive cylinder strength

Cylindrical attenuation factor for CAM

Iyy Moment of inertia of a section in y-y direction

Length of the beam span

Diagonal mass matrix

MDOF Multi-degree-of-freedom

Mw Moment magnitude of earthquakes

Axial compression force at balance failure (3D RUAUMOKO)

Axial compression yield force (3D RUAUMOKO)

s Spacing of reinforcing bars

Spectra velocity SDOF Single-degree-of-freedom

u&&g Ground acceleration as a function of time

vf Shear capacity of FRP wrap (MPa)

Source factor for CAM Factor for equivalent rectangular block of concrete in compression

Inelastic attenuation factor for CAM Factor for equivalent rectangular block of concrete in compression

Factor used in CAM Factor for equivalent rectangular block of concrete in compression

ρsc Reinforcement ratio

CHAPTER 1 INTRODUCTION

1.1 Background

Because Singapore is located on a stable part of the Eurasian Plate, with the

nearest earthquake fault 400 km away in Sumatra, the buildings in Singapore were

designed according to British Standard (BS8110 1985) without any seismic provision

Since BS8110 code does not consider the earthquake loads, buildings in Singapore can

be referred to as gravity-load-designed (GLD) structures That is they are designed to

resist gravity load, wind load and a notional lateral load (1.5% of the weight of the

building), without any earthquake loads However, due to the far-field effects of

earthquakes in Sumatra (Balendra et al 1990), buildings in Singapore, of which most

are reinforced concrete (RC) shear wall and frame structures, are occasionally

subjected to tremors that occur at the Sumatra fault and the subduction of the

Indian-Australian plate into the Eurasian Plate (as shown in Figure 1.1)

In the last two years (2004 and 2005), tremors were felt several times in

Singapore due to the strong earthquakes at Sumatra, according to the reports on

newspaper (Strait Times, Today and Lian He Zao Bao) These earthquakes are: the

Singapore), the earthquake on March 28, 2005 (Mw=8.7, 600 km away), the earthquake

on Dec 26, 2004 (Mw=9.3, 950 km away), the earthquake on July 25, 2004 (Mw=7.3,

magnitude Mw is used instead of the more widely-known Richter scale ML, because it

Seismic waves generated at the epicenter can be categorized into two groups: (1)

high frequency waves, which have high intensity but damp out rapidly during

propagation; (2) low frequency waves, which possess large displacement properties

and damp out relatively slowly Because the high frequency waves generated at

Sumatra faults damp out quickly during propagation to Singapore, the seismic waves

reach Singapore bedrock are often rich in low frequency waves Although the peak

ground acceleration (PGA) of the low frequency waves may be very low, the induced

motions may have disproportional high displacement and possibly high velocity

characteristics In addition, such low frequency waves may be amplified more than 10

times to reach the ground level through the soft soil layers in Singapore, if the natural

period of the soil site is close to the predominant period of the bedrock motion (This

kind of amplification due to resonance is called site amplification effect) Furthermore,

such amplification may be further enlarged, if the natural period of the building

supported on such soil sites is close to the predominant period of the ground motion

Therefore, due to the large displacement properties that low frequency waves possess

and the amplification by the soil, buildings in Singapore may suffer from large

displacement that may cause damage Balendra et al (1990) revealed that an

earthquake of magnitude 7.6 at Richter scale 400 km away from Singapore could cause

a shear force demand of 2-4% of weight of the building, which is larger than the

notional lateral load(1.5% of weight) in the design

Thus, there is a need to evaluate the seismic vulnerability of RC shear wall and

frame structures in Singapore in case a larger or nearer earthquake may occur in the

future Experimental and numerical studies of seismic vulnerability of such structures

should be carried out, and seismic retrofitting scheme should be proposed if needed

1.2 Literature review

1.2.1 Overview of seismic studies of RC GLD structures

Besides buildings in Singapore, old RC structures built in other non-seismic or

low-seismic regions, like the eastern and central United States and Korea, are also

GLD structures Such RC GLD structures are the result of the old codes which did not

consider seismic load Although different codes are used, RC GLD structures in

different regions share some common features such as: (Aycardi et al 1994; Lee and

Woo 2002b)

1 Minimal transverse reinforcement in columns or shear walls for confinement and

shear resistance; the spacing of hoops of column/shear wall transverse

reinforcement is relatively large

2 Little or no transverse shear reinforcement in beam-column joints

3 Lap splices of columns or shear walls are located in the potential plastic hinge

zones at the bottom

The concern for the vulnerability of RC GLD structures is relatively recent, and

research work has been carried out in USA, Korea and Singapore According to

different design codes, the research work can be divided into three groups:

1 Research work at SUNY Buffalo University, Cornell University, UC Berkeley and

Hanyang University on the seismic performance and retrofitting technology of RC

GLD frames designed according to American code (ACI nonseimic code) (Aycardi

et al 1994; Bracci et al 1995b, 1995a; Kunnath et al 1995b, 1995a; ElAttar et al

1997; Mosalam and Naito 2002; Han et al 2004)

2 Research work at Korea University on the performance of RC GLD frames

designed according to Korean practice of nonseismic detailing under seismic

loading (Lee and Sung-W 1998; Lee and Woo 2002b, 2002a)

3 Research work in Singapore on the performance and retrofitting scheme of GLD

RC frames and shear walls designed according to British Standard (BS8110 1985)

(Balendra et al 1999; Balendra et al 2001; Koh 2003; Kong et al 2003a; Perry

2003; Suyanthi 2003; Li et al 2004; Dhakal et al 2005)

1.2.2 Research of GLD buildings designed according to ACI code

The research work at SUNY Buffalo University started from the component

tests and analysis, followed by the test of scaled frames This led to the development of

a seismic evaluation methodology, and finally an effective and economical retrofitting

scheme

Aycardi et al.(1994) tested four column specimens, loaded with low and high

levels of axial forces, with and without lap splices, representing interior and exterior

columns at floor slab and beam soffit levels They also tested one 1/3 scale external

beam-column joint subassemblage and one 1/3 scale internal joint subassemblage

They reported that structural components with detrimental details of GLD structures

could reach their flexural strength and still sustain their gravity loads during large

cyclic deformations For example, column specimens were found to be able to sustain

at least 70% of the maximum load capacity for two cycles of 4% drift It was reported

that the failure mode was dependent on the level of axial load, although the failure in

the columns was flexurally dominated For instance, the failure mode for exterior

subassemblage was weak beam-strong column mechanism while the interior

subassemblage was weak column-strong beam mechanism In this study, it was also

reported that the plastic hinge lengths in columns were 0.74h to 1.35h under large axial

load and 0.47h to 0.6h under low axial load, where h is the column dimension in the

direction of bending

Based on the work of Aycardi et al.(1994), Bracci et al (1995a) performed a

3-story 1/3 reduced scale model test on the shake table They concluded that the

behavior and may suffer from side-sway/soft-story collapse mechanism at the ultimate

load It was found that a critical mechanism formed at a base shear of 15% of the

building weight

The evaluation methodology including damage modeling proposed by Kunnath

et al (1995b) was used to identify the seismic performance of 3-, 6-, and 9-story GLD

RC buildings The analysis of tests was performed by a computer program IDARC

(First version was introduced in 1987, and the latest version was distributed in 2004),

which is a nonlinear analysis program for frame-wall structures subjected to seismic

excitations Results of the analysis matched well with those of the experiments

The research of Cornell University was to see the performance of the GLD

building under realistic seismic force ElAttar et al.(1997) tested a 1/6 scale 2-story

office building and a 1/8 scale 3-story one-bay by three-bays office buildings on the

Cornell University shake table They concluded that the reinforcement details in the

GLD building may not be sufficient to develop a complete failure mechanism, because

before these details fail, a premature soft-story mechanism may occur due to the lack

of sufficient strength in columns as compared to beams In addition, they drew the

conclusion that the GLD frame without infilled walls would experience large lateral

deformations during the test and the slab contribution to beam negative moment

flexural strength may alter the relatively ductile strong column-weak beam mechanism

to soft-story mechanism It was noted in their study that significant stiffness

deterioration occurred in the test due to wide cracking and pullout of some reinforcing

bars This stiffness deterioration resulted in large displacements and a pronounced

P-Δ effect

The research work in UC Berkeley concentrated on the seismic performance of

the GLD perforated waffle-slab systems In this study, Mosalam and Naito (2002)

conducted two reduced scale identical specimen, modeling half the height of an

interior column with a portion of the waffle slab bounded by the centerlines of adjacent

bays In the test, bidirectional quasi-static lateral loading is applied at the test column

end It was found that the system possessed high deformation ductility with only minor

damage to the waffle slab The failure mode reported was the formation of stable

plastic hinging at the junction between the column and the waffle-slab joints

The study of Han et al.(2004) focused on the seismic performance of Ordinary

Moment-Resisting Concrete Frames (OMRCF) designed primarily for gravity loads A

3-story OMRCF was designed according to the minimum design requirements in ACI

code (ACI318 1989) A 1/3-scale 3-story model was constructed and tested under

quasi-static reversed cyclic lateral loading It was found that the overall behavior was

quite stable without abrupt strength degradation It was reported that the base shear of

15% of the weight was achieved in the test Capacity spectra method was carried out to

evaluate the seismic performance of the frame, and the results showed that this 3-story

detailing

The research work of Korea University focused on the seismic performance of

low-rise RC frames designed to the Korean practice of non-seismic detailing, and the

influence of masonry infills and scale effects of such RC frames

Lee and Sung (1998) investigated the influence of the scale factor to the seismic

performance of RC frames They manufactured one 1/2.5 and one 1/10 model

subassemblages, and applied quasi-static reversed load to these subassemblages The

strength, stiffness, energy dissipation, failure modes and local deformations were

compared It was concluded that the strength and crack pattern of the two models were

similar, while the initial stiffness, energy dissipation capacity and failure mode of the

two models were different The initial stiffness and energy dissipation capacity of the

1/10 model were smaller than those of 1/2.5 model

Lee and Woo (2002a) performed 2-bay 3-story 1/5 scale masonry-infilled RC

frame tests, under a series of earthquake simulation loading and a pushover loading, to

investigate the influence of masonry-infills They concluded that masonry infills could

be beneficial to the seismic performance of the structure, because they contribute to the

large increase in the stiffness and strength of the global structure, and the amount of

the increase in strength is greater than additional inertia forces due to the infills

Lee and Woo (2002b) conducted a 1/5 scale 3-story RC frame designed

according to the Korean practice of non-seismic detailing, and tested the model using

shake table A pushover test was finally performed, due to the limitation of the

capacity of the shake table It was reported that the model revealed fairly good

resistance to the higher levels of earthquakes, though it was not designed against

earthquakes The drifts observed were approximately within the allowable limit The

analysis by IDARC-2D program (First version was introduced in 1987, and the latest

version was distributed in 2004) revealed that the overall displacement ductility ratio

was 2.4 and the overstrength coefficient was 8.7

The above research is for GLD buildings designed according to ACI code

(ACI318 1989) and Korean practice of nonseismic detailing (Lee and Woo 2002b),

which are different from British Standard (BS8110 1985) used in Singapore Since

different design codes or practices may result in different detailing, which will

influence the seismic behavior of the structures, the studies in USA and Korea could

only be used for reference if the structures in Singapore are under consideration In the

next section, the research work in Singapore is reviewed

1.2.4 Research of GLD buildings in Singapore designed according to

BS8110 code

Both experimental and numerical studies have been carried out in Singapore, for

the GLD buildings designed according to BS8110(1985)

(2004) conducted a test of a scaled model, which is the lower critical region of a

25-story RC shear wall structure The results of the tests showed that RC shear wall

has significant overstrength but nominal ductility

At the Nanyang Technological University, Dhakal et al (2005) performed six

full-scale tests of beam-column joint specimens under reversed cyclic loading The

cyclic displacements were applied at different speeds varying from slow quasi-static

loading to high-speed dynamic loading (20HZ), to evaluate the response of the

specimens subjected to excitations of different frequencies It was found that the joints

experienced severe damage, and joint shear failure occurred before the formation of a

plastic hinge in the adjoining members

In the above research, monotonic static pushover loading or cyclic loading were

used in the pseudo-static loading tests However, no comparison between pushover

behavior and cyclic behavior has been carried out, so whether the results from

pushover test can be a simplified representation of the cyclic behaviors is yet to be

found out

For the seismically designed structures, it was found that pushover behavior can

be a backbone representation of the cyclic behavior Lefas et al (1990) and Penelis

and Kappos (1997) tested some rectangular walls, under both pushover and cyclic

loads They found that the tested pushover curves formed the backbone of the tested

cyclic curves Dolsek and Fajfar(2002) performed the pushover analysis of a 3D

3-story frame specimen, which had been tested under cyclic loading The resulting

analytical pushover curves enveloped the tested cyclic curves Ei-Tawil and Kuenzli

(2002) analyzed the shear walls, which were tested by Oesterle et al (1976,1979) and

Pilakoutas and Elnashai (1995), under both pushover and cyclic loading They found

that the analytical pushover curves could be the backbone of the analytical cyclic

curves

However, for the GLD structures, because the details such as lap splices and

reinforced confinement at joints, are quite different from the structures with seismic

provision, whether the pushover loading test behavior will form the backbone

representation of the cyclic behavior is yet to be established For this purpose, tests of

GLD frame and shear wall structures under pushover and cyclic loading, and the

comparison between the pushover and cyclic results should be carried out

In conclusion, the above research in Singapore mainly focuses on experimental

studies under pushover loading or cyclic loading The comparison of the experimental

results between pushover loading and cyclic loading is yet to be carried out

Furthermore, since the scaled model tests might not represent the behavior of a full

scale building, FEA numerical models for full scale buildings should be established

overstrength and considerable ductility due to the redistribution of internal forces in

the inelastic range The influence of infill walls on the response modification factor

was also investigated in their research Balendra et al.(2001) performed nonlinear

pushover analysis using ABAQUS to investigate a 16-story RC GLD frame building

Kong (2004) established a microscopic FEA model using ABAQUS to calculate the

capacity of a 25-story GLD frame-shear wall building

In the above numerical studies, pushover analysis using ABAQUS was

performed However, there are some limitations of ABAUQS, such as it cannot be

used to perform dynamic collapse analysis, and it cannot consider the initial effective

stiffness reduction due to the micro cracks in the members Therefore another software

RUAUMOKO (Carr 2002a, 2002b) was introduced recently, and FEA models using

RUAUMOKO were developed by Koh (2003), Perry (2003) and Suyanthi (2003) In

their studies, full scale FEA models for a 16-story frame building and a 25-story GLD

shear wall-frame building were established Seismic capacities of these buildings were

calculated by performing the pushover and dynamic collapse analysis Their research

provided a good analytical tool to evaluate the seismic performance of the full scale

buildings However, some important parameters in their models, like initial effective

stiffness reduction factors and parameters for hysteresis rules, were obtained from the

research work of buildings designed according to other codes instead of BS8110(1985)

code The accuracy of their FEA models needs to be evaluated, and those important

parameters should be obtained from the experimental test of GLD buildings designed

according to BS8110(1985) In the following sections, the literature review of such

parameters is addressed

Initial stiffness

Definition of initial stiffness is based on the effective stiffness definition given

by Paulay and Priestley (1992) A typical nonlinear relationship between loads and

displacement under pushover loading is shown in Figure 1.2 For simplification, an

idealized bilinear response is often used in analysis to represent the actual observed response In Figure 1.2, S defines the yield or ideal strength y S of the member, and i

y

∆ defines the corresponding yield displacement The slope of the idealized linear

elastic response,K =S y/∆y, is defined as initial stiffness, which is equal to the effective secant stiffness of the real load-displacement curve at a load of about 0.75S y

It was reported that the initial stiffness obtained from the test was less than that of

the uncracked gross section (Bracci et al 1995b; Filiatrault et al 1998) According to

Bracci, the reason for such differences stems from the initial cracking of the member

sections due to gravity loads (particularly in beams), micro-cracking generated from

concrete curing, and minor construction loadings Macgregor (2005) pointed out that

micro-cracks, which consist of bond cracks and mortar cracks, have an enormous

effect on stiffness of structures Aktan et al.(1985) also observed that the

initial stiffness is less than the theoretical stiffness of the uncracked gross section are

also reported by many other researchers (Hirosawa 1975; Kenneally and Burns 1986;

Elnashai et al 1988, 1990; Pilakoutas and Elnashai 1995; Lopes 2001)

In order to incorporate the reduction effect of initial stiffness into the analysis,

the reduced initial stiffness instead of the gross section stiffness should be used in the

FEA model As an example, Bracci (1995a) used the initial stiffness reduction factors

obtained from the test, to establish FEA model using IDARC-2D (First version was

introduced in 1987, and the latest version was distributed in 2004) It was shown that

the FEA model could predict the experimental results well Filiatrault et al.(1998)

conducted FEA modeling using RUAUMOKO, and validated the model using the

experimental results obtained from shake table tests of two half-scale RC moment resisting frames In the FEA model, they used the equivalent moment of inertiaI (less eq

than half of I , the moment of inertia of gross section) obtained from the test, as the g

input of initial moment of inertia As a result, good correlation between numerical and

experimental results was achieved Harries et al (1998) carried out nonlinear dynamic

analyses of four couple wall prototype structures using RUAUMOKO In their FEA

models, reduced initial stiffness, based on the Explanatory Notes on Canadian

Standards Association (CAS) Standard A23.3 Clause 21.2 (CPCA 1995), was used as

input parameters Han et al (2004) performed modal analysis of 3-story OMRCF

structure using SAP2000(1997) In the FEA models, they used the reduced stiffness as

specified in the ATC-40 document (1996), as a result, the 1st modal natural period predicted by FEA is only 4.3% larger than the test result Therefore, in order to

establish an accurate FEA model, it is wise to select the appropriate initial stiffness

reduction factor, which is the ratio of the initial stiffness (or equivalent moment of inertiaI ) obtained from the test over the uncracked section stiffness (or gross eq

moment of inertiaI ) g

Such a reduction factor has been suggested in some literature Bracci (1995a)

suggested 0.55 to 0.6 for columns and 0.25 to 0.35 for beams Filiatrault et al.(1998)

suggested 0.40 for columns and 0.38 for beams of the nominally ductile frame, and

0.43 for columns and 0.34 for beams of the ductile frame Harries et al (1998) used 0.7

for columns and 0.17-0.2 for coupling beams The reduction factors recommended by

Paulay and Priestley (1992),CPCA (1995) and ATC-40 (1996) are listed in Tables 1.1-

1.3

Hysteresis rules

Otani (1981) compared several most commonly used hysteresis rules and

discussed the effect of the stiffness parameters to the structural response He pointed

out that stiffness would reduce due to flexural cracking of concrete and tensile yielding

of steel, in the process of loading Thus, in order to capture the seismic behavior of

Q-HYST hysteresis rule were obtained from the shake table tests Harries et al (1998)

selected the Modified Takeda degrading stiffness hysteresis rule (Otani 1981) to model

the RC coupling beams, and chose a Ramberg-Osgood hysteresis rule to model the

steel coupling beams The parameters for hysteresis rules were obtained by comparing

the predicted responses with test results

According to Otani (1981), for RC structures, the most often used hysteresis rule

is the Modified Takeda model The diagram of this hysteresis rule and the

corresponding parameters are shown in Figure 1.3 Three parameters are used to define

this hysteresis rule They are:

1 Unloading stiffness parameter,

j

j m u

d

d d k

k

))(

1( 0 − −

=

fatness of the hysteresis loop and the plastic residual deformation

2 Post yield stiffness factor or bilinear factor,

0

1

k

k

r = This parameter influences the

strength enhancement after yielding

3 Reloading stiffness parameter,

degradation between two nearby cycles